- This topic has 38 replies, 7 voices, and was last updated 4 years, 9 months ago by
.
-
Posts
-
“normal” behaviour is easy to achieve: rightclick on loader logo, profile associations, untick “use personal settings”, the tick use personal settings” –> normal colors.
Interesting. Double-click to reload calibration doesn’t help I presume?
It would be good to have the actual video card gamma table contents while the colors are messed up. You can use the Curve Viewer for this.
My problem has also been with the combination of Radeon+1903+DisplayCAL. No problem at all before 1903, no problem if I use a Nvidia card, no problem if I uninstall DisplayCal.
No need to uninstall anything. Either enable Windows internal calibration loader (limited to 8-bit precision though, and has a scaling bug) in Windows color management advanced settings (implicitly disables DisplayCAL profile loader doing any calibration loading) or right-click profile loader tray icon -> remove check mark from “Load calibration at login and preserve calibration state”.
Haven’t reported to the Radeon team, as my feeling has always been that its’ related to the bug I oringally mentioned, about the conflict with the Windows profile loader. I’ve always thought, and though I misquoted the bug and the patch you provided, what I’m trying to say is that the error only occurs when DCProfile Loader loads a profile. And any correction to it occurs when DCProfile Loader loads a profile.
This morning, I followed the workaround about disabling the MS Task for profile loading, and the problem has disappeared completely. So, it may be related to AMD. But it seems to be a conflict between 1903’s profile loader and AMD/DisplayCAL. It seems to me that MS is a better place to start than anywhere else.
Still waiting for the videoLUT contents while the problem has occured.
Currently it is not clear at all where the problem actually happens on a technical level. I’m thinking it must be related somehow to what color format (pixel encoding) the graphics card is sending to the monitor (e.g. if the graphics card sends YCbCr, but the monitor expects to receive RGB). This is likely triggered by a Windows 1903 related graphics driver anomaly in the Radeon drivers, otherwise it would occur with different brands as well.
I followed the workaround about disabling the MS Task for profile loading, and the problem has disappeared completely.
Note that the DisplayCAL profile loader will try to disable that task automatically when run (which is a work-around for the 1903 banding problem which is not graphics card specific, and requires logging out and back in after disabling to become effective – see sticky thread in this forum). Do you see the respective entries in DisplayCAL-apply-profiles.log (these will be there even if the task is already disabled)?
Hopefully, someone can provide that. I only chimed in on the previous thread because the issue seemed to have been dismissed and closed as user-error, and I wanted to make sure someone knew that it was not.
re: Logs. Haven’t looked, not sure where that would be, and have already spent all the time on this that I can. I can tell you that the tasks in question were not disabled, regardless of logging out and back in (maybe that’s normal, I don’t know). As I mentioned, disabling the tasks manually has resolved the issue for me. Thanks.
Still waiting for the videoLUT contents while the problem has occured.
todays situation: tinted monitorcolors happen only when monitor awakes from sleep, not after a boot.
the moment i enter curve viewers menu, colors revert to normal. means i didn’t catch any data to submit.
regards!
Radeon RX-560, last driver (19.7.2), W 10 pro 18362.239.
Still waiting for the videoLUT contents while the problem has occured.
todays situation: tinted monitorcolors happen only when monitor awakes from sleep, not after a boot.
the moment i enter curve viewers menu, colors revert to normal. means i didn’t catch any data to submit.
regards!
Radeon RX-560, last driver (19.7.2), W 10 pro 18362.239.
I solved the problem by clean up the AMD driver with AMD Cleanup Utility
and restore default windows driver. Never happen again!
@kevin-4 use the standalone Curve Viewer (inside DisplayCAL folder in Windows start menu). When the problem occurs, tick the “show calibration from video card” checkbox.
Checking the pixel format in the AMD Catalyst Control Center might also be a good idea.
Got it. And done. Also: Had ZERO troubles since my last post, but today have had 3 interruptions:
- When Profile Loader loaded at startup
- Both times my monitor woke from sleep
re: Pixel Format, I’m using DisplayPort so AMD says that isn’t relevant. as the setting only applies to HDMI connections. I’ve included a screenshot of that setting just in case, though .
Calibration looks sane in your case. What’s more interesting is that only some desktop elements seem affected by the messed up colors, i.e. the curve viewer window looks completely unaffected (except titlebar). What’s also interesting is that the problem appears in screenshots, so (as I suspected) it is not calibration related.
re: Pixel Format, I’m using DisplayPort so AMD says that isn’t relevant. as the setting only applies to HDMI connections
DisplayPort does support RGB as well as YCbCr, so the setting should be relevant in the same way it is for HDMI.
Had ZERO troubles since my last post, but today have had 3 interruptions:
When Profile Loader loaded at startup
Both times my monitor woke from sleepSome questions:
- What happens when you change pixel format? Does it fix the problem?
- Does quitting the profile loader fix the problem? In which case I’d be interested in the logs (C:\Users\<Your Username>\DisplayCAL\logs\DisplayCAL-apply-profiles*.log)
- When you re-launch the profile loader from the start menu, does the problem re-occur?
- When you disable calibration loading via the profile loader’s right-click menu (icon turns gray), does the problem still occur?
- Does specifying a delay for the task (in Task Scheduler, right-click DisplayCAL Profile Loader Launcher, select Properties -> Triggers -> click on “On Login” -> Edit -> Under “Delay for ” enter e.g. 1 Minute)
Ok, I actually have been able to go through those just now:
- Changing the pixel format fixes the issue for a second, and then when the profile loader re-loads it immediately reintroduces it.
- No, quitting it does not fix it.
- Yes, 90% of the time. Literally, out of ten tries, 9 of them produced this and 1 did not.
- Disabling the profile loader (or quitting it) requires that I turn the monitor off and on again in order to get back to a normal screen. After that, if I re-enable calibration loading, the problem did not reoccur.
- I’ve added the delay and will report back later on if that makes a difference over the next few restarts.
re: screenshots. I am not sure what happened there. I have definitely taken screenshots where the problem shows up. However, today, the screenshots don’t capture the problem.
Changing the pixel format fixes the issue for a second, and then when the profile loader re-loads it immediately reintroduces it.
Interesting, as the calibration (the actual values) don’t seem to reflect that. Maybe the AMD driver caches the values internally so even if the correct values get set by SetDeviceGammaRamp, it uses messed up cached values internally.
No, quitting it does not fix it.
Good in a sense, anything else would have surprised me.
Disabling the profile loader (or quitting it) requires that I turn the monitor off and on again in order to get back to a normal screen. After that, if I re-enable calibration loading, the problem did not reoccur.
My interest was more if with a permanently disabled profile loader (only via right-click menu -> remove checkmark from “load & preserve calibration”, no other means) the problem still occurs or not. If it does, we would be back to square one. If not, it’s probably due to some weird side-effect of SetDeviceGammaRamp with the AMD driver and 1903 (although the technical details would still be unclear, i.e why it seems to work via Windows internal calibration loader unless MS would have changed that to no longer use SetDeviceGammaRamp).
re: screenshots. I am not sure what happened there. I have definitely taken screenshots where the problem shows up. However, today, the screenshots don’t capture the problem.
Oh so the curve viewer screenshot above actually has pink title bars as part of Windows theme and is not due to the problem?
